Cancer in nickel-processing workers in New Caledonia.

tautology. Clearly, an independent measure is critical so that the question "what is the effect of attractiveness on RS?" can be asked, and hypotheses about function can be tested. Thissen and Martin's suggested "free-choice" design, wherein birds could interact freely, has not proved feasible because the intense intrasexual competition generated swamps the choice process. While not all aspects of this methodological problem have been worked out, we have data (1, 12) that attractiveness can be measured prior to pair formation without physical contact between participants. The finding that "attractive" females are less successful reproductively than "intermediate" females is intriguing. One possible reason for this result has been discussed (3). Attractiveness is only one of several components affecting RS in polygynous and promiscuous species (13); its role in monogamous species is hardly understood. Understanding its importance requires information on the effect of attractiveness on several aspects of reproduction, including timing, frequency of breeding, eagerness to breed, the "willingness" of individuals to invest in their offspring (14), and the behavioral complementarity of various pair combinations. NANCY BURLEY Department ofEcology, Ethology and Evolution, University of Illinois, Urbana 61801 References and Notes